Review Reports - The Needs of Women Using Homelessness Services: The Results of Collaborative Research in London

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Materials and methods:

"The methodology centred on gaining insight into the view, experiences and patterns of service use..." This paragraph does not describe the methodology used in the study. What follows is a description of methods; you need to either expand on the methodology, or only focus on methods.

More detail needs to be given as to how the data was analysed and whether this was a deductive process only (and why). More clarification needs be given as to the purpose of the survey, given the limitations you highlight.

There is little indication of how you analysed the qualitative data, other than a brief reference to thematic analysis. I could not find an indication of the quantity of data you had (eg average length of interview) or how many service providers you interviewed, and who they were. There is no indication of the limitations of using online interviews only, and the safeguarding measures put in place for the women, if they were affected by the interviews.

There is no indication as to how the four datasets were integrated in a mixed methods design - it seems to be parallel design but this isn't mentioned or reflected in the analysis, ie the datasets seem to have been analysed in isolation of each other, with little apparent effort at integration or drawing rounded inferences.

Results:

It would have been good to see a comment on the likely representativeness of the survey sample, and a further commentary on the limitations of using administrative data emanating from provider services.

I would have liked to see some critical commentary on the longitudinal tracking of service use; for example the sample suggests that only half of the participants experienced sexual violence. This is (unfortunately) surprisingly low. This dataset is potentially really powerful in highlighting instances where these women could have been better supported by services. A detailed analysis should highlight the pathways in and out of homelessness that these women have experienced; this would be closely aligned to the call for this special edition. Instead, you present the data as prevalence, which seems to underplay the potential of the longitudinal nature of your data. For example, at what average age did the women enter in contact with services? Which services? How many contacts have they had on average and how spaced out were they? etc...

It would have been good to see more detail on which services you are referring to; you talk about 'agencies working with women who were experiencing homelessness' without stipulating which these are. At times it looks like you are referring to homelessness services only, but you also mention child removal (are these court proceedings data?) domestic abuse and addiction, which is confusing. Greater clarity as to what the data entails is needed.

The findings are peppered with broader literature that do not fit neatly there. For example, there is a whole paragraph on the history of homelessness services in the UK, which seems to bear little relation to the findings. This should either be in background or in the discussion, to free up some space to discuss the actual findings. Similarly, some of the findings around contacts with child protection system for example feature some general data that could be referred to in the introduction of in the discussion.

The quotes need some kind of participant reference. The qualitative data seems to only present two quotes; one not being marked as such and two of them appearing to be coming from service providers... Meaning that this paper presents one quote from nine interviews with these incredibly vulnerable women. There is no indication of the overall themes that have emerged from that data. Did you analyse all the qualitative data as one dataset, or two? Were there themes in common between the women and service providers? Any contradictory or moderating findings between the two? There was great potential here to add granularity and explanations to your quantitative data, which is left unexplored. The process of analysis needs to be more transparent, and the results section needs to focus on your findings only. Much more could be made of the analytical integration of the four datasets.

Discussion:

The discussion reads like a manifesto for Housing First. While your assertions are probably right, this is a research paper and your research findings need to be discussed as such. A mention of solutions is fine, but there is no mention of the strengths and limitations of the study (ie online interviews with marginalised and vulnerable women; PIT...) or how the findings add to the broader research literature, which is paramount.

Overall, the paper is very descriptive and highlights in various points where it concurrs with pre-existing research, making it difficult to see its unique contribution. The data, in particular the longitudinal service use data, is really quite unique, but the analysis does not do it justice. You note on a couple of occasions that this isn't a statistically representative sample, but you could have instead sought out rich descriptions and highlight stronger patterns in the data, given the longitudinality of the dataset. Statistical significance should be irrelevant here. There is no indication of the quantity of data that you have, and how it presented itself, or, again, how you integrated the various datastrands in the analysis.

This is such an important field of research and your data is potentially really powerful. However the paper does not demonstrate the kind of rigorous and transparent analysis that is needed to do justice to those women, and influence practitioners and policy makers.

Author Response

We thank the referees for their comments and the time they took to review the paper.
We are unable to fully address all the comments made by the referees because they took very divergent views of the paper.
Some comments on language and syntax we are not able to address because it appears some of the referees did not have (British) English as a first language - we are unable to Americanise the English - and/or confused some elements of the paper with others, e.g. interpreting quotations as being part of the text of the paper.
We have added some detail to the methods section following comments and some additional subheadings in response to referee's comments, but just to note, this is not a methodology journal and it is important to balance this against being able to contextualise and discuss the results.
A small additional section has been added on the analysis as some reviewers wanted more information on how mixed methods are used in this sort of social research.
The paper is not a 'manifesto for Housing First'. The bulk of evidence on Housing First, including multiple RCTs, does strongly indicate it is more effective than earlier service models. However, the broader points here do really centre on the need for services that understand and respond to women’s needs, which is explained at some length.
We do not agree with the statements that the paper fails to represent women’s experiences of homelessness accurately, or that it fails to add anything new to analysis. This is a much more extensive and detailed analysis than much of the work hitherto conducted in this space.

Reviewer 2 Report

Comments and Suggestions for Authors

Please see attached file.

Comments for author File: Comments.pdf

Author Response

We thank the referees for their comments and the time they took to review the paper.
We are unable to fully address all the comments made by the referees because they took very divergent views of the paper.
We have simplified and expanded the methods section in response to comments from reviewers who found some parts of it unclear or wanted more detail.
We have added some detail to the methods section following comments and some additional subheadings in response to referee's comments, but just to note, this is not a methodology journal and it is important to balance this against being able to contextualise and discuss the results.

Reviewer 3 Report

Comments and Suggestions for Authors

This is research about an important issue in current society and yes until relatively recently there was a lack of understanding about women and homelessness but that has improved considerably over the last decade or so.

A journal article should provide a ‘new’ way of examining a topic or providing new evidence. As such the abstract should present a problem statement, the main argument of the paper and a summary of the evidence provided by the research – a summary of the journal article.

This abstract does not do this. Essentially the abstract states that this research supports the findings of previous research without indicating what this article offers that is different and this follows through the whole article.

The introduction does not clearly set up the argument of the paper – it presents no aims, objectives or research questions and therefore there is no reasoning for the methodology, the findings or the discussion even though some good points are made. This means the article raises many questions in the readers mind.

I am sure the data collected can be turned into a journal article, the author(s) just need to find the angle they want to discuss. I am assuming there were aims and objectives in the original research and I suggest the author(s) return to these and do some reading about writing journal articles or get someone to help them.

I am concerned that the abstract states the research was conducted by a team then line 300 states ‘I think’

Comments on the Quality of English Language

The quality of the English in the paper is generally good however at numerous times what are presented as sentences are not sentences. For example:

Line 161 the sentence … Interviews with stakeholders from five organisations at the core of provision for 160 women experiencing homelessness in and around the London Borough of Camden. should say…..Borough of Camden were conducted.

Line 230 the sentence International data have repeatedly suggested low numbers of women living rough and in basic emergency shelters, while numbers of women increase significantly if supported housing and people experiencing hidden homelessness (Baptista, 2010) is missing an ending like ‘are included.’

Line 236 the sentence By the mid 1990s, smaller scale hostels, frequently using individual bedrooms rather than shared sleeping areas, alongside increased use of what were called resettlement and floating support services that supported people directly into settled housing, which were similar to (though less intensive than) Housing First services ( Quilgars, 2000).

Line 300-306 A very very long rambling sentence

Author Response

We are unable to fully address all the comments made by the referees because they took very divergent views of the paper.
Some comments around language and syntax we are not able to address because it appears some of the referees did not have (British) English as a first language - we are unable to Americanise the English - and/or confused some elements of the paper with others, e.g. interpreting quotations as being part of the text of the paper.
We are unable to amend the results and conclusions to come into line with the expectations of some of the referees, the findings are the findings and the interpretation and analysis is that of the authors.
We are unable to address those comments that a) are grounded in images and ideas about women’s experience of homelessness that differ from what this and other UK (and European) evidence suggests and b) dislike the paper and its presentation but are not actually highlighting clear limitations with the research, analysis or presentation. Again, the referee reports are also inconsistent in their appraisal and comments on the paper

Reviewer 4 Report

Comments and Suggestions for Authors

The authors of the manuscript "The Needs of Women Using Homelessness Services: The Results of Collaborative Research in London" (socsci-2934095) use a mixed methods approach to study variation in the experiences of homelessness among women from multiple perspectives and data sources. The study speaks to emerging research and theory on hidden homelessness as well as practical service delivery implications. Although I like much about the paper, the strengths of the manuscript and its potential are dampened by several weaknesses that will require substantial revisions to the content, structure, and organization of several key sections of the paper.

Abstract: Is the academic team based in London as well (in addition to the homelessness service sector collaborators)? Rather than saying collected data “on”, use “with”. Use mixed methods as the framing for the study type. Include the sample size for each data source. No need to state that participants agreed to interviews in the abstract. Would sustained and recurring homelessness fit with what is described as chronic homelessness in the literature?

Introduction: Bring in more context on subgroups of homelessness (see Lee, Tyler, and Wright 2010 below) to lead to more emergent literature on hidden homelessness. Several recommended references include the following:

Crawley, J., Kane, D., Atkinson-Plato, L., Hamilton, M., Dobson, K., & Watson, J. (2013). Needs of the hidden homeless – no longer hidden: a pilot study. Public Health, 127(7), 674–680. https://doi.org/10.1016/j.puhe.2013.04.006

Krabbe, J., Jiao, S., Guta, A., Slemon, A., Cameron, A. A., & Bungay, V. (2021). Exploring the operationalisation and implementation of outreach in community settings with hard-to-reach and hidden populations: protocol for a scoping review. BMJ OPEN, 11(2). https://doi.org/10.1136/bmjopen-2020-039451

Lee, B. A., Tyler, K. A., & Wright, J. D. (2010). The New Homelessness Revisited. Annual Review of Sociology, 36(1), 501–521. https://doi.org/10.1146/annurev-soc-070308-115940

Richard, M. K., Dworkin, J., Rule, K. G., Farooqui, S., Glendening, Z., & Carlson, S. (2022). Quantifying Doubled-Up Homelessness: Presenting a New Measure Using U.S. Census Microdata. Housing Policy Debate, 1–22. https://doi.org/10.1080/10511482.2021.1981976

Watson, J., Crawley, J., & Kane, D. (2016). Social exclusion, health and hidden homelessness. Public Health, 139, 96–102. https://doi.org/10.1016/j.puhe.2016.05.017

On page 2, line 49 there is an extra semi colon: (; Baptista et al, 2017).

At the end of the introduction or as a section before the methods, more should be stated about the collaboration between the services and academic team. More should also be stated about how participants were recruited. There should also be a clearer research question or objective before the methods section.

Materials and Methods: Frame and organize this study as mixed methods. Describe each data source (survey, service history, and interview) for women affected by homelessness in terms of nesting. As I understand it, there is a likelihood that there is overlap, i.e., nesting, but the level is not known due to the precautions to protect the identities of participants. Then describe the sampling for key informant interviews. Describe how the data sources can be used to triangulate results and strengthen overall findings and study implications. A good paper to reference is below:

Small, M. L. (2011). How to Conduct a Mixed Methods Study: Recent Trends in a Rapidly Growing Literature. Annual Review of Sociology, 37(1), 57–86. https://doi.org/10.1146/annurev.soc.012809.102657

The limitations of nonrandom sampling and sample size should be moved to a limitations section in the discussion. Remove mention of challenges, e.g., page 3 line 145 “Orchestrating and conducting the semi-structured interviews with women experiencing homelessness presented some challenges to the research team.” Challenges with doing the interviews remotely, rather than in person, should be moved to a limitations section in the discussion as well.

In general, organize by data source, describe the sampling methods, and protocols/schedules for collecting data, and provide example questions. Headings to separate the data sources would be helpful.

Results:

Organize results by data source with headings. Consistently use percentages rather than counts. For example, change: “Twenty-three of the women were recorded as having been made homeless as a result of domestic abuse” to the percentage.

Language about sampling limitations and representativeness should go into the limitations section in the results, e.g., page 5 line 207: “While the sample was not large enough to be statistically representative…”

Use more data from the 9 interviews. I think there should be at least 3 quotes. Within the quotes, clean the sentences to avoid typos, page 6, line 268 (mould to mold) and to be read more easily, e.g., page 6, line 269 “I lived rough for a year, some, a year, but then in a hostel”, could be “I lived rough for a year…but then in a hostel”. Also, though I’ll leave it to the discretion of the authors and editors, I’d recommend including the actual expletives that were stated, rather than censuring them.

After the key informant interviews, describe the overall findings drawn from the data sources in aggregate.

Discussion. Revise this section to include overall takeaways, limitations, implications for future research, policy, and practice, and a short conclusion. Much of what is written already can fit in this structure. Limitations can be moved from the language I’ve noted in previous sections to this new section. The discussion should include more about hidden homelessness.

Thanks for the opportunity to review this paper. I appreciate the research and wish the authors all the best with their continued work.

Comments on the Quality of English Language

I have no major concerns with the writing. A few small things are to make sure to clean up the participant quotations in which there were several typos. Also, avoid unneeded repetition in the paper. I have attempted to help with this in my general comments. The word centred is probably overused in the manuscript as well. It appears at times this could be replaced with focused on, engaged with, comprised, or included.

Author Response

We thank the referees for their comments and the time they took to review the paper.
We are unable to fully address all the comments made by the referees because they took very divergent views of the paper.
There are suggestions to add references which are not directly relevant to this paper, i.e. generic papers and work that is largely focused on male experiences of homelessness. As this literature does not directly apply to the subject of this paper, we are unable to incorporate this suggestion.
We have simplified and expanded the methods section in response to comments from reviewers who found some parts of it unclear or wanted more detail.
We have added some detail to the methods section following comments and some additional subheadings in response to referee's comments, but just to note, this is not a methodology journal and it is important to balance this against being able to contextualise and discuss the results.
We are unable to address those comments that a) are grounded in images and ideas about women’s experience of homelessness that differ from what this and other UK (and European) evidence suggests and b) dislike the paper and its presentation but are not actually highlighting clear limitations with the research, analysis or presentation. Again, the referee reports are also inconsistent in their appraisal and comments on the paper. Some comments suggested a lack of familiarity with the subject and current debates within it.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Thank you for providing some methods clarifications; these are imperative in reporting research rigorously and transparently. The comment on length of interviews is welcome, though your previous experience of length of interview would have been better replaced by references of peer reviewed articles, focusing on online vs face to face interviews. Citing other research with homeless women and the length of their interviews would have been helpful too; 25 mins is very short.

My query about methods of analysis was not to hear about how mixed methods are used generically, but how they were used in this research, in particular. It would have been better if you could have detailed your analytical framework. SPSS, Excel and NVivo are not methods of analysis. The fact that encrypted recorders were used should be in data collection methods, not analysis.

I maintain my comment about the history of homelessness taking valuable space in the findings section. It is not a finding of this research.

My point about Housing First was that the discussion did not draw directly enough from the findings; I am not questioning Housing First, but merely pointed out that you present it as a solution when your findings in and of themselves, do now allow you to do that, or at least not to the extent that it takes much of the space in the discussion. There could have been more and better use of broader research literature. I maintain my point about the fact that it doesn't feel like you have done justice to the longitudinal data, or made a clear and explicit case for the contribution to knowledge that this study makes (other than you saying it does, in the response).

Overall, you have paid lip service to some of my comments, and ignored others. I take exception at your comment on referees not having English as first language.

Author Response

Thank you for your comments. We are afraid that we do not agree that 25 minutes on average is ‘very short’. We have acknowledged that the interviews tended to be shorter than has been the case with face to face work and we have added a couple of references to that effect as suggested. The interviews, as stated, mostly took place over the telephone due to restrictions placed on research staff and often the participants were walking around / sat on the streets during the interview and maintaining engagement was tricky to say the least. I am sure it is not necessary for me to add this here, but accessing and maintaining interviews with vulnerable and or marginalised populations, and attempting to do this remotely, is not especially straightforward.

There is a lot of discussion of methods in here and this is a paper with a word limit, which follows the expected format for this journal – if we had a separate subheading for ‘data collection’ it would only be a sentence long. The analytical framework is thematic grid analysis, which works on an open basis and has grounded theory at its core, i.e. there is no preconception of what the data are or how they will be interpreted, inside the analysis is led by the data, within a wider practice of enabling research participants to prioritise and list their own concerns, experiences and interpretations rather than imposing a predetermined set of factors on respondents. We have added a little material on this.

I maintain my comment about the history of homelessness taking valuable space in the findings section. It is not a finding of this research.

As was explained before, this sets the context for the research, explains the situation in which women were becoming homeless and experiencing various trajectories through homelessness. This was not a comment made by other referees who seem to have accepted this was an intrinsic and necessary part of the paper. It remains.

We’d have to say this is more a question of personal perception and particular views of what women’s homelessness is, rather than actually being a comment we can adequately respond to at this point. As to the comment that we have not made extensive use of the wider literature, we have included the literature that we feel is appropriate, and we should also note that we have actually written quite a lot of that wider literature but were asked to remove it by the editors due to strict limits on self-citation. We can’t adjust our data, our analysis or our interpretation of the data to match what the referee expects from a paper on women’s homelessness.

Overall, you have paid lip service to some of my comments, and ignored others. I take exception at your comment on referees not having English as first language.

Yes, we have not made changes that we believe would be to the detriment of the paper. Peer review is a process that is designed to create a dialogue around improving academic work, it is not a set of instructions from one reviewer that are supposed to be simply obeyed without question (and indeed were we to do so it would contradict the suggestions from the other three reviewers). The comment about language was not directed at anything from this referee, but from another who commented on the poor grammar in sections of the paper which were actually the quotes from women experiencing homelessness and whose attempts to correct syntax and grammar strongly suggested that they were not working in their first language.

Reviewer 2 Report

Comments and Suggestions for Authors

The revision of this manuscript addresses several concerns; however, key issues remain. While it is appreciated that the authors tried to address reviewer comments, they seemed to miss attention to key points. What is the research question? (The methods used for this research centered on gaining insight into the views, experiences and patterns of service use among women experiencing homelessness, line 101-102). Are they really the first to look at women’s homelessness in the UK or the Camden area? Have they tried to do too much but in fact, have confused the data findings? It is not clear what they are trying to shed light on.

The authors should explain how their use of both anonymous survey and the anonymized tracking survey were handled. The explanation provided continues to lead to questions. The authors say that they could not rule out double counting of respondents or if the timeframes from which they are drawing their results from match the current situation (line 160). However, in the results they present information as it were two distinct samples (line 209) vs. (line 230+).

Highlighting 3 quotes, and 2 stakeholder comments from such a small sample to define a finding of lack of coordination of services and why women had various sleeping arrangements could be more robust to add to the voices of the respondents—which would then lead to the authors being able to answer their research goal which they seem to do on lines 372-375. While they offer HFW as a solution, in their own words, they recognize that this may not happen any time soon (lines 403-405). Their work, therefore, does not provide immediate assistance or recommendations for doable actions—in the meantime—or this neglected population.

In addition, typically, the results would present the results, and all the other discussion items would be in the discussion section. Perhaps this manuscript will be clearer if the service survey data was removed.

The authors seem to disagree with many of the comments offered to them to enhance the manuscript.

Author Response

Thank you for your comments. This is an academic journal article, it does not have a ‘research question’ because it explores data on a complex form of social exclusion with reference to the existing literature, theory and through a mixed methods approach. There’s no suggestion we are the first to look at women’s homelessness in the UK or in Camden, but this study was one of the largest and most comprehensive attempts that there has so far been in the UK to understand women’s homelessness. We are not at all confused by the study or the findings and are unfortunately unable to address this comment from the referee.

We do explain what the data were and how they were collected. No, we do not present the data as two distinct samples, they are the results of two elements of a mixed methods approach and we have, as the referee has noted, said that there may have been some overlap (although controls were present around double counting).

The first comment here is not clear and thus we cannot address it. As for the second point regarding ‘doable actions’, we are unaware of any requirement for single academic papers to present complete policy solutions to complex social phenomena.

We cannot agree with this, and it is also not in line with comments from the other referees.

The authors seem to disagree with many of the comments offered to them to enhance the manuscript.

This is true in instances where we do not believe the comments / suggestions would enhance the paper or where reviewer’s comments are conflicting. Some suggestions however have been very helpful and have been included.

Reviewer 3 Report

Comments and Suggestions for Authors

The changes made to the paper have improved its readability and clarified the objectives and findings of the research.

It is much clearer now (though at times still obtuse) that the right questions need to be asked and the right measurement tools used to establish the 'true' level of a phenomenon and that in the case of women experiencing homelessness and this impacts the responses to an issue. These are important points to make and worthy of publication.

There are some minor changes still needed that would again improve the quality of the paper for the reader:

Final sentence of Introduction section states : “This paper draws on the results of a mixed methods piece of research carried out in collaboration with programmes and services working with women experiencing homelessness in London, during the summer of 2021” – but still the authors do not define what they are researching – the intersecting errors they've spent more than half of the introduction outlining or the circumstances of women’s homelessness in London. It appears to be the later so state this!

Lines 102-103 in materials and methods provides the aim of the project now: gaining insights into the views, experiences and patterns of service use among women experiencing homelessness. If these lines were added to the last sentence above in the introduction the reason for the paper is much clearer.

Still need to outline the purpose of each data collection in adding to the paper's argument in the methods section and again in the results section. For example , the anonymised survey was not intended to enumerate scale/nature of homelessness – but what was its purpose?; Administrative data -why collected? Authors need to just state the value and purpose of each data collection and how that adds to the argument they want to substantiate.

Results section outlines the findings from each individual data collection but doesn’t summarise what each brings to the results collectively – need to tie together otherwise article could just have reported on the findings of one or two of the data collections as suggested by one of the reviewers.

Again reference in the discussion needs to be made back to the considerable discussion in the Introduction with regard to the cluster of intersecting data errors, otherwise the words spent on this in the introduction should be reduced and used in the discussion as one explanation for the findings.

Grammatically change line 187-188 just join the first two sentences:

Interviews with five stakeholders from five organisations at the core of provision for women experiencing homelessness in and around the London Borough of Camden. These interview centred on stakeholder perceptions of women’s homelessness, the nature of needs and strengths and limits of existing service coordination and provision.

Author Response

The changes made to the paper have improved its readability and clarified the objectives and findings of the research.

Thank you for your comments and suggestions on this.

There are some minor changes still needed that would again improve the quality of the paper for the reader:

We have spelled things out further in a new couple of sentences at the start of the introduction.

We’ve hopefully addressed this in the response to the preceding comment.

This was a multi-method study, multiple methods cross-check one another and add a depth not present when reliant on a single data source, such as a survey, we do think we’ve illustrated that and explained it sufficiently.

That is not really how mixed methods analysis works from our experience and unfortunately we cannot really address that again otherwise we believe it would make the paper convoluted and difficult to read.

Apologies we don’t understand this. What are ‘intersecting data errors’?

Grammatically change line 187-188 just join the first two sentences:

We’ve fixed this, thank you.

Reviewer 4 Report

Comments and Suggestions for Authors

To begin, I’d like to commend the authors for making several improvements that have made the manuscript easier to follow. This was the reason for my previous suggestions to use headers and signal which data sources are being referred to (not to turn the paper into a methodological paper). Second, I had not meant to Americanize the paper in my first review. I apologize for that being the message received (in combination with the other reviews, I suspect). However, I still believe it important to consider sociological literature that has shown changes and reasons for the differences in the composition of people affected by homelessness over time (or changing attention to/realization that homelessness is a broader experience)—by age, gender, and race, etc. There is also a growing body of literature on hidden homelessness--as a type of homelessness that contrasts with stereotypical images. I think this paper could be in dialogue with more scholarship on hidden homelessness. Yes, this literature is from the US. Does this literature completely overlap with this study? Nope. That's why it's potentially very compelling. That said, attending to this would only require a few additional sentences. Although I think this would serve the paper well, I will leave it to the editor and authors to make this decision.

Author Response

Thank you for your comments and previous suggestions. American conceptualisations (and indeed definitions and experiences) of homelessness do differ quite markedly from those in the UK and North Western Europe, there are also very significant differences between those societies, which are still welfare states in the European sense and the way in which social, housing, welfare, public health and other policies work in America. The authors were amongst those who first demonstrated how trajectories women took through homelessness were more likely to involve hidden homelessness in comparison to men and have also contributed to the literature that conceptualises and discusses the measurement of hidden homelessness in a UK and European context (and we do mention hidden homelessness 18 times and make pretty extensive reference to the - admittedly UK and European – literature on this). References related to some of this work were removed due to limits on self-citation however.