nestedAlpha: Compute Alpha for Nested Data Structures

In kthorstmann/horst: Functions for the processing of questionnaire data

Description

This function returns a nested alpha for multilevel data as proposed by Nezlek (2016).

Usage

nestedAlpha(item.level.1, level.2, level.3, data, ...)

Arguments

item.level.1	The variable that includes the items on level 1.
level.2	The variable that includes the nests on level 2, for example days or measurement occasions in which the items are nested.
level.3	The cluster in which the days are clustered, usually persons, groups, classes.
data	The data frame that contains all of the above variables.
...	Other parameters that can be passed on to lmer.

Details

Nezlek, 2016, describes one way in that the internal consistency as a measure for reliabilty can be obtained for longitudinal data. The data structure for these types of examinations usually is items (for example all items of one scale) nested in days (for example 10 days of a diary study), which are again nested in persons (all subjects).

To estimate the reliability, one needs to consider the nested structure.

Please consider the following points before running the analysis:

• Coding of items As is also the case with Alpha, all items must be coded in the direction of the construct.

• All items are weighted equally

• All items are assumed to be parallel (i. e. equal error-variances, equal loadings, and unidimensional)

• The variable on level-1 should not be standardized or transformed, but the raw scores

The function relies on the lmer function in order to run the multilevel model. The multilevel model is run with standard settings, however, these can be changed using the ...-argument.

Value

An alpha for nested data.

Author(s)

K. T. Horstmann [e-mail]

References

Nezlek, J. B. (in press). A practical guide to understanding reliability in studies of within-person variability. Journal of Research in Personality. http://dx.doi.org/10.1016/j.jrp.2016.06.020.

Examples

sample data frame, 10 persons, 10 days with 4 items per day
person <- sort(rep(letters[1:10], 40))
day <- rep(sort(rep(seq(1:10), 4)), 10)
variable <- rnorm(400, 50, 10)
data <- data.frame(person, day, variable, stringsAsFactors = FALSE)

nestedAlpha(item.level.1 = "variable",
            level.2      = "day",
            level.3      = "person",
            data         = data)

kthorstmann/horst documentation built on May 20, 2019, 7:04 p.m.