Frequently Asked Questions
1. For what tests is INNDI currently collecting data?
Over time, INNDI will aggregate normative data for dozens of cognitive tests, questionnaires, and other diagnostic instruments. Our current priorities include the tests listed below. In order to derive regression-based norms, we need the actual test scores of individual test-takers. We cannot use summary statistics, such as mean or percentile scores for groups of participants.
[one_half_first]
- Mini-Mental State Examination (MMSE)
- Verbal Fluency
- Digit Span
- Trail Making Test
- Wisconsin Card Sorting Test
[/one_half_first]
[one_half]
- Stroop Color-Word Test
- Clock Drawing Test
- Rey-Osterrieth Complex Figure Test
- Rey Auditory Verbal Learning Test
- Raven’s Progressive Matrices
[/one_half]
2. Are there inclusion and exclusion criteria?
Yes, but they are guidelines rather than rules. Because INNDI databases include data collected by different investigators using varied sampling methods, normative samples differ in their precise makeup. Our goal is to pool representative samples of generally healthy adults and children. We do not want samples that consist only of “super healthy” persons because the resulting regression-based norms will not generalize well to the broader population. Therefore, we specify relatively few exclusion criteria. Neither do we want samples that include persons with severe health problems that are known to impair cognition. When potential contributors have questions about which participants are suitable for INNDI, we discuss their questions and help contributors select participants for inclusion. If information about diagnoses and treatment is available, typical exclusion criteria include dementia, other neurological illnesses/conditions (such as Parkinson’s disease, epilepsy, multiple sclerosis, and a history of stroke or severe traumatic brain injury), severe mental illnesses (such as schizophrenia or affective psychosis), alcohol or drug addiction, or treatment with anti-psychotic, anti-convulsant, or anti-Parkinsonian medications.
3. How do you control the quality of data that comprise INNDI normative databases?
While the quality of sub-samples that comprise any given INNDI normative database will vary, we maintain the strictest possible quality control through several means. Normative data must meet certain criteria to be eligible for inclusion. The most important of these is that we include test data only from studies that have been described in a peer-reviewed publication or presentation. This does not mean that the normative must be published, but rather that the methods used to collect the data have been peer-reviewed. In addition, we check data rigorously for out-of-range values and serious distributional departures from expectation based on other samples. Finally, whenever possible, we compare the results of our statistical analyses of data sent to INNDI with those published by the contributor.
4. How should data be formatted for submission to INNDI?
Data should be sent in an electronic spreadsheet format, such as an Excel spreadsheet or SPSS, Stata, or SAS data file. Note: We cannot use aggregated data, such as tables of mean scores or percentiles. The actual test scores of individual participants are required. In addition, we request as many of the demographic variables listed below as possible. Finally, we prefer but do not require that variables be labeled in English.
In addition to each participant’s test scores, contributors are asked to provide data for as many of the following demographic variables as possible:
[one_half_first]
- Age (years)
- Sex (male/female)
- Ethnicity or race
- Years or level of education
[/one_half_first]
[one_half]
- Literacy (yes/no)
- Test version (if applicable)
- Primary language spoken
- Country of test administration
[/one_half]
5. Will INNDI pay for data preparation or transfer?
No. Unfortunately, we do not have funding to pay for data preparation or transfer.
6. How will the data be analyzed?
For each test, INNDI begins by pooling normative data from different countries into a central database. Then, based on a modification of the statistical methods described here, we will derive regression-based norms that enable users to demographically-adjust raw scores for any test-taker whose demographic characteristics are represented in the database. Thus, INNDI databases provide the empirical basis to determine whether a Mini-Mental State Exam (MMSE) score of 24/30 is abnormal for a highly educated, 60-year-old woman from France, but normal for an illiterate, 79-year-old man from Uruguay. In this way, regression-based norms provide raw score-to-percentile conversions that are tailored to an individual’s unique characteristics.
7. How will INNDI norms for each test be disseminated?
As soon as regression-based norms are developed for each test, an article describing the INNDI normative sample and look-up tables based on the regression-based norms will be published in a peer-reviewed scientific journal. Each INNDI article will include tables that allow the reader to look up raw score-to-percentile conversions for male and female test-takers across the age span that is represented in the database. Also, adjustments for a person’s educational level and the country and language in which testing was conducted will be provided in ancillary look-up tables. In this way, regression-based norms for each test will be disseminated as a free global resource.
8. Are collaborators acknowledged for their contribution to INNDI?
Yes. One or two contributors from each site that contributes data will be acknowledged by name in publications to which they contribute data. In addition, each person who contributes data to an INNDI normative database is identified in several places on this website. Finally, hyperlinks to the websites of many contributors and multi-site studies appear throughout these pages.
9. Will contributors or other users have direct access to INNDI databases?
No. Contributors agree to share normative data with INNDI, but not with others. Therefore, it would violate the terms of our agreement with contributors to release data or allow others direct access to INNDI databases. However, our goal is to make using the regression-based normative data as widely available as possible.
10. Can contributors co-author publications of INNDI norms?
We welcome varying levels of participation by those who contribute data to INNDI. Contributors who provide data for 5% or more of the normative sample for a test will be invited to co-author the article in which norms for that test are published. There may be circumstances under which investigators who contribute fewer cases will be invited to co-author an article. This determination will be made on a test-by-test basis.