Evaluation 2020

Prevalence of Site-Level Missing Data in a National Evaluation of Programs for At-Risk Families 

8 days ago

Data can be missing at the individual level or at a higher level, such as family, classroom, or program site.  Increased knowledge of higher-level missing data is necessary to develop evaluation design and statistical methods to address it.  We aim to identify the prevalence and patterns of site-level missing data in a national dataset of programs serving at-risk families. Data represent self-report demographic and program participation measures from 17 programs for at-risk families with a total 1,420 adult participants. Nearly one third of all sites had site-level missing data, including 10 sites from 5 programs.  Site- or program-level missing data accounted for 27.07% of all missing data. The high prevalence of site-level missing data suggests an urgent need to limit the occurrence of missing data (e.g., by generating buy-in at the site level) and to refine statistical methods to account for it when it occurs.  Evaluators should generate buy-in at the site, as well as program, level. 

Statistics
0 Favorited
14 Views
1 Files
0 Shares
11 Downloads

Related Entries and Links

No Related Resource entered.

Tags and Keywords

Attachment(s)
pdf file
Prevalence of Site-Level Missing Data in a National Evalu...   2.66MB   1 version
Uploaded - 10-14-2020
Data can be missing at the individual level or at a higher level, such as family, classroom, or program site. Increased knowledge of higher-level missing data is necessary to develop evaluation design and statistical methods to address it. We aim to identify the prevalence and patterns of site-level missing data in a national dataset of programs serving at-risk families. Data represent self-report demographic and program participation measures from 17 programs for at-risk families with a total 1,420 adult participants. Nearly one third of all sites had site-level missing data, including 10 sites from 5 programs. Site- or program-level missing data accounted for 27.07% of all missing data. The high prevalence of site-level missing data suggests an urgent need to limit the occurrence of missing data (e.g., by generating buy-in at the site level) and to refine statistical methods to account for it when it occurs. Evaluators should generate buy-in at the site, as well as program, level.