CFA with repeated measures per individual, and independent data set

[Álvaro Josué Vega Hidalgo]

Hello,

I am trying to make a CFA in r with lavaan from data of three independent experiments, made to the same 15 individuals. Each experiment was repeated to the same individual eight times.

One of these experiments got 5 observed variables, and the other two experiments have one variable measured per each. Therefore, I do not have just one dataset like HolzingerSwineford1939 does. Actually, I have three datasets, which look like this:

Experiment 1 Dataset:

Individual        var1        var2        var3        var4        var5
Ind1                23          2                1            3            32

Ind1                19          10              3            3           39

.

.

(Eight repetitions)

.

.

Ind2                32        33                55        54             32

Ind2                21        33                94        43             23

.

(Eight repetitions)

.

.

Ind3 ... (15 individuals with 8 repetitions)

Experiment 2 Dataset

Individual                Var6

Ind1                        100

Ind1                           99

ind1                          89

.

.

(Eight repetitions)

Ind2                           44

Ind2                          34

.
.

(Eight repetitions)

Ind3... 15 individuals with 8 repetitions

Experiment 3 Dataset

Individual                Var7

Ind1                        100

Ind1                         99

ind1                         89

.

.

(Eight repetitions)

Ind2                        44

Ind2                        34

.
.

(Eight repetitions)

Ind3... 15 individuals with 8 repetitions

The problem is that I do not know how to write any model to fit, given that I have three datasets, and repeated measures per individual.
One of the models that I would run is like the following:

HS.model <- ' latent1 =~ var1 + var2 + var3 + var4 +var5     

latent2 =~ var6 

latent3 =~ var7 '

I hope you can help me. I really need to make this analysis.

[Terrence Jorgensen]

15 individuals. Each experiment was repeated to the same individual eight times.

This is a lot of observations on very few people.  The first thing that occurs to me would be to use some kind of Dynamic P Technique model, where the "population" is an individual person, and all their observations are stacked in a single column.  Three other columns would indicate the experiment (1-3), variable (1-7), and replication (1-8), and another column would indicate, as usual, which person (1-15). The 15 individuals could be treated as 15 groups in a multigroup model, so each "group" is a single person's sample of observations.

But I have no idea what your actual research goal is, so I don't know whether this is a useful approach.  This isn't even a lavaan issue, so you should instead post this question on SEMNET:

www2.gsu.edu/~mkteer/semnet.html

Terrence D. Jorgensen

Postdoctoral Researcher, Methods and Statistics

Research Institute for Child Development and Education, the University of Amsterdam

UvA web page: http://www.uva.nl/profile/t.d.jorgensen