getResults | R Documentation |
First the IRT model should be defined using defineModel
. Afterwards,
call runModel
with the argument returned by defineModel
to start the estimation.
The last step then is to create a results frame using getResults
.
getResults( runModelObj, overwrite = FALSE, Q3 = TRUE, q3theta = c("pv", "wle", "eap"),
q3MinObs = 0, q3MinType = c("singleObs", "marginalSum"), omitFit = FALSE,
omitRegr = FALSE, omitWle = FALSE, omitPV = FALSE, abs.dif.bound = 0.6,
sig.dif.bound = 0.3, p.value = 0.9, nplausible = NULL, ntheta = 2000,
normal.approx = FALSE, samp.regr = FALSE, theta.model=FALSE, np.adj=8,
group = NULL, beta_groups = TRUE, level = .95, n.iter = 1000,
n.burnin = 500, adj_MH = .5, adj_change_MH = .05, refresh_MH = 50,
accrate_bound_MH = c(.45, .55), sample_integers=FALSE, theta_init=NULL,
print_iter = 20, verbose = TRUE, calc_ic=TRUE, omitUntil = 1, seed=NA)
runModelObj |
The object returned by |
overwrite |
Logical. Should result files be overwritten if exist? |
Q3 |
Logical. Estimate the Q3 statistic according to Yen (1984)? Note: this is only
possible for uni-dimensional models. If |
q3theta |
Specify whether the Q3 statistic should be estimated using PVs, WLEs or EAPs as the theta variable. |
q3MinObs |
Q3 statistic might be untrustworthy if item covariance estimation is based on very few observations. Define the minimum number of observation which should be fulfilled for Q3 estimation. |
q3MinType |
If |
omitFit |
Logical. Should item fit values be included into the results? |
omitRegr |
Logical. Should regression parameters and their standard errors be included into the results? |
omitWle |
Logical. Should WLE estimates be included into the results? |
omitPV |
Logical. Should plausible values be included into the results? |
abs.dif.bound |
Applies only if DIF analyses are performed before. When DIF-Parameter are evaluated, this specifies the critical value for absolute DIF. See the details section for further information. |
sig.dif.bound |
Applies only if DIF analyses are performed before. When DIF-Parameter are evaluated, this specifies the critical value for confidence interval DIF. See the details section for further information. |
p.value |
Applies only if DIF analyses are performed before. When DIF-Parameter are evaluated, this specifies the critical p-value for confidence interval DIF. See the details section for further information. |
nplausible |
Applies only if |
ntheta |
Applies only if |
normal.approx |
Applies only if |
samp.regr |
Applies only if |
theta.model |
Applies only if |
np.adj |
Applies only if |
group |
Applies only if |
beta_groups |
Applies only if |
level |
Applies only if |
n.iter |
Applies only if |
n.burnin |
Applies only if |
adj_MH |
Applies only if |
adj_change_MH |
Applies only if |
refresh_MH |
Applies only if |
accrate_bound_MH |
Applies only if |
sample_integers |
Applies only if |
theta_init |
Applies only if |
print_iter |
Applies only if |
verbose |
Applies only if |
calc_ic |
Applies only if |
omitUntil |
Argument is passed to |
seed |
Fixed simulation seed. This value is directly passed on to the |
If defineModel
was run with software Conquest, a path argument ('dir'
)
is necessary. The path argument is optional for software TAM. If 'dir'
was
specified, getResults
additionally writes its output into the specified folder,
using the analysis.name
argument for file naming. Otherwise, getResults
only returnes the result data frame.
If DIF analyses were performed before, the user can specify the criteria according to
which DIF should be interpreted or evaluated. By default, the ETS criteria (Zieky, 1993)
are used which classify DIF into three distinct categories, "A", "B", or "C". Small DIF ("A")
corresponds to absolute DIF values below .43 (no significance test is performed here); medium
DIF ("B") corresponds to absolute DIF values between .43 and .64 which are significantly
different from zero. High DIF ("C") corresponds to absolute DIF values above .64 which are
significantly different from .43 (DeMars, 2011; Monahan et al. 2007). Alternatively, the
three arguments abs.dif.bound
, sig.dif.bound
, and p.value
allow to
specify user-defined dichotomous criteria. If items should be flagged as DIF, if the
absolute value increases 0.5 and significantly exceeds 0.1 at a alpha level of 0.05, use
abs.dif.bound = 0.5
and sig.dif.bound = 0.1
and p.value = 0.95
.
A data frame in the long format with ten columns.
model |
The name of the model (as specified by the user in |
source |
The estimation software (i.e, conquest or TAM) |
var1 |
The variable name for which the corresponding value is given, i.e. its indicator. |
var2 |
Additional variable information if necessary. |
type |
Type of coefficient (for example, random or fixed). |
indicator.group |
The type of the group the corresponding variable belongs to. |
group |
The group the corresponding variable belongs to. Note: group is nested within |
par |
The type of the parameter. |
derived.par |
Optionally: The derived parameter. |
value |
The value of the corresponding estimate. |
DeMars, C. E. (2011). An analytic comparison of effect sizes for differential item functioning. Applied Measurement in Education, 24 (3), 189-209. https://doi.org/10.1080/08957347.2011.580255
Monahan, P. O., McHorney, C. A., Stump, T. E. & Perkins, A. J. (2007). Odds ratio, delta, ETS classification, and standardization measures of DIF magnitude for binary logistic regression. Journal of Educational and Behavioral Statistics, 32 (1), 92-109. https://doi.org/10.3102/1076998606298035
Zieky, M. (1993). Practical questions in the use of DIF statistics in item development. In P. W. Holland & H. Wainer (Eds.), Differential item functioning (pp. 337-347). Hillsdale, NJ: Lawrence Erlbaum.
# see examples in the help file of defineModel()
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.