Work Description

Title: Survey Distributional Methods Analyses Data Open Access Deposited

h
Attribute Value
Methodology
  • The data used for this study comes from the survey responses of a study on Mobility-on-Demand in 2019. From the survey responses, we gathered the income information of the participants and the method through which they were recruited. From the study, we gathered the cost associated with each survey distributional method, and data such as number of people contacted, number of people responded, and number of complete/valid responses. We calculated a series of rates, including response rates, cost effectiveness, and average income through simple Excel functions, such as sums, additions, divisions, and means.
Description
  • We compared the response rates, cost, and the average income of participants pertaining to 6 different survey distribution methods used in an initial study about mobility-on-demand services. We used the data to identify survey and recruitment methods that are more effective in reaching hard-to-reach populations. All the raw data used for calculations and the calculations themselves can be found in the attached spreadsheets.

  • Initial analyses have identified in-person onsite recruitment as one of the better methods of reaching hard-to-reach populations, and is calling for continued work in improving research methods in the field of HCI.
Creator
Depositor
  • xuecong@umich.edu
Contact information
Discipline
Keyword
Citations to related material
  • Yan, X., Zhao, X., Han, Y., and Hentenryck, P. V. (2019). Mobility-on-demand versus fixed-route transit systems: an evaluation of traveler preferences in low-income communities. https://poverty.umich.edu/files/2019/02/Yan_et_al_WorkingPaper_Preference_for_mobility_on_demand.pdf
  • Atkinson, R., and Flint, J. Accessing Hidden and Hard-to-Reach Populations: Snowball Research Strategies. "Social Research Update" 33 (Jan 2001).
  • Buranyi, S. Rise of the racist robots: how ai is learning all our worst impulses, Aug 2017. Retrieved June 11, 2019 from https://www.theguardian.com/inequality/2017/aug/08/rise-of-the-racist-robots-how-ai-is-learning-all-our-worst-impulses.
  • Dillahunt, T. R., Erete, S., Galusca, R., Israni, A., Nacu, D., and Sengers, P. Reflections on design methods for underserved communities. In Companion of the 2017 ACM Conference on Computer Supported Cooperative Work and Social Computing (New York, NY, USA, 2017), CSCW ’17 Companion, ACM, pp. 409–413.
  • Erete, S., Israni, A., and Dillahunt, T. An intersectional approach to designing in the margins. Interactions 25, 3 (Apr. 2018), 66–69.
  • Foster, A. Concerning issue with driverless cars, Mar 2019. Retrieved June 11, 2019 from https://www.news.com.au/technology/innovation/motoring/on-the-road/driverless-cars-may-be-more-likely-to-hit-darkskinned-people-study-finds/news-story/b19959d01ef865f15bb336275b8903e8.
  • Johnston, L. G., and Sabin, K. Sampling hard-to-reach populations with respondent driven sampling. Methodological Innovations Online 5, 2 (aug 2010), 38.1–48.
  • Macaulay, A. C., Commanda, L. E., Freeman, W. L., Gibson, N., McCabe, M. L., Robbins, C. M., and Twohig, P. L. Participatory research maximises community and lay involvement. BMJ 319, 7212 (sep 1999), 774–778.
  • Maestre, J. F., Eikey, E. V., Warner, M., Yarosh, S., Pater, J., Jacobs, M., Marcu, G., and Shih, P. C. Conducting research with stigmatized populations: Practices, challenges, and lessons learned. In Companion of the 2018 ACM Conference on Computer Supported Cooperative Work and Social Computing (2018), ACM, pp. 385–392.
  • Paterson, J. M., and Maker, Y. Why does artificial intelligence discriminate?, Jun 2019. Retrieved June 11, 2019 from https://pursuit.unimelb.edu.au/articles/why-does-artificial-intelligence-discriminate.
  • Strohmayer, A., Laing, M., and Comber, R. Technologies and social justice outcomes in sex work charities: fighting stigma, saving lives. In Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems (2017), ACM, pp. 3352–3364.
  • Sydor, A. Conducting research into hidden or hard-to-reach populations. Nurse researcher 20, 3 (2013).
Resource type
Curation notes
  • On August 21, 2019, metadata was updated with additional creators and citations to related material.
Last modified
  • 08/21/2019
Published
  • 07/16/2019
DOI
  • https://doi.org/10.7302/0qnz-fw92
License
To Cite this Work:
Xu, X., Yan, X., Dillahunt, T. R. (2019). Survey Distributional Methods Analyses Data [Data set], University of Michigan - Deep Blue Data. https://doi.org/10.7302/0qnz-fw92

Relationships

This work is not a member of any user collections.

Files (Count: 12; Size: 344 KB)

Thumbnailthumbnail-column Title Original Upload Last Modified File Size Access Actions
Deep_Blue_Data_Xu_Readme.txt 2019-07-01 2019-07-15 3.66 KB Open Access
00_data_calculations.xlsx 2019-07-15 2019-07-15 10.8 KB Open Access
00_data_end_values.csv 2019-07-15 2019-07-15 1.28 KB Open Access
01_individual_income_raw.xlsx 2019-07-15 2019-07-15 20.6 KB Open Access
02_average_income_calculations.xlsx 2019-07-15 2019-07-15 78 KB Open Access
02_average_income_end_values.csv 2019-07-15 2019-07-15 860 Bytes Open Access
03_inperson_cost_calculations_ca....xlsx 2019-07-15 2019-07-15 73.6 KB Open Access
03_inperson_cost_calculations_en...s.csv 2019-07-15 2019-07-15 1.57 KB Open Access
04_additional_cost_table.xlsx 2019-07-15 2019-07-15 9.19 KB Open Access
05_overall_results_calculations.xlsx 2019-07-15 2019-07-15 68.1 KB Open Access
05_overall_results_end_values.csv 2019-07-15 2019-07-15 573 Bytes Open Access
06_responserates_cost_table.xlsx 2019-07-15 2019-07-15 75.5 KB Open Access

Date: 16 July, 2019

Dataset Title: Survey Distributional Methods Analyses Data

Dataset Creators: X. Xu, J. Yan, T.R. Dillahunt

Dataset Contact: Xuecong Xu xuecong@umich.edu

Funding: Poverty Solutions

Key Points:
- We compared the response rates, cost, and the average income of participants pertaining to 6 different survey distribution methods used in an initial study about mobility-on-demand services (Yan et al., 2019).
- We used the data to identify survey and recruitment methods that are more effective in reaching hard-to-reach populations.

Research Overview:
This work examines the survey recruitment methods used in a past study to contribute to the continued efforts of involving disadvantaged populations to combat AI discrimination. Initial analyses have identified in-person onsite recruitment as one of the better methods of reaching hard-to-reach populations, and is calling for continued work in improving research methods in the field of HCI.

Methodology:
The data used for this study comes from the survey responses of a study on Mobility-on-Demand by Yan et al. in 2019. From the survey responses, we gathered the income information of the participants and the method through which they were recruited. From the researchers of Yan's study, we gathered the distributional cost associated with each survey distributional method, and other data such as number of people contacted, number of people responded, and number of complete/valid responses. With these data, we calculated a series of rates, including response rates, cost effectiveness, and average income through simple Excel functions, such as sums, additions, divisions, and means. The calculations are then compared for interpretation of the data.

Because we only used the income data from the original survey, the complete survey responses are not included in this repository. To get more information on the original survey, please see Yan's study: https://poverty.umich.edu/files/2019/02/Yan_et_al_WorkingPaper_Preference_for_mobility_on_demand.pdf.

Instrument and/or Software specifications: NA

Files contained here (where there are two instances of a numbered file, the filename ending in "calculations" is an Excel sheet including calculations used in the research, while the filename ending with "end_values" indicates a .csv with final calculated values only):

00_data: this spreadsheet contains miscellaneous data associated with each distributional method, including the number of people reached/contacted, the number of complete/valid response, and the distributional cost.

01_individual_income_raw: this file contains all the self-reported household income of the survey respondents organized by distributional methods. This data is used to calculate the average income in each group presented in file 02_average_income.

02_average_income: this file contains a table that list the number of participants in each income group organized by distributional methods. The "average" function in Excel is used to calculate the average in each group.

03_inperson_cost_calculations: this file shows the process of calculating the cost associated with in-person onsite distribution method. Specifically, the 2nd and 3rd tables towards the end have the work records of two research assistants. The 1st table shows the calculations of the cost paid to the RAs given their individual hourly wages.

04_additional_cost_table: this table summarizes the additional cost (total cost minus compensation to the survey respondents) of this distributional method with explanation.

05_overall_results: this table contains data on the number of people contacted, complete response, valid response, and cost. Excel calculations including additions and divisions are performed.

06_responserates_cost_table: this table summarizes % of valid responses received and cost per response associated with each distributional method.

Related publication(s):
Yan, X., Zhao, X., Han, Y., and Hentenryck, P. V. (2019). Mobility-on-demand versus fixed-route transit systems: an evaluation of traveler preferences in low-income communities. https://poverty.umich.edu/files/2019/02/Yan_et_al_WorkingPaper_Preference_for_mobility_on_demand.pdf

Use and Access:
This data set is made available under a Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)

To Cite Data:
Xuecong Xu, Tawanna Dillahunt, and Xiang Yan. (2018). Reaching Hard-To-Reach Populations: an Analysis of Survey Recruitment Methods [Data set]. University of Michigan - Deep Blue. https://doi:10.7302/0qnz-fw92

Download All Files (To download individual files, select them in the “Files” panel above)

Best for data sets < 3 GB. Downloads all files plus metadata into a zip file.



Best for data sets > 3 GB. Globus is the platform Deep Blue Data uses to make large data sets available.   More about Globus

Remediation of Harmful Language

The University of Michigan Library aims to describe library materials in a way that respects the people and communities who create, use, and are represented in our collections. Report harmful or offensive language in catalog records, finding aids, or elsewhere in our collections anonymously through our metadata feedback form. More information at Remediation of Harmful Language.