I designed a scale to measure ideological views toward curriculum. I developed seven questions, each asking participants about their value orientation on one component of curriculum development. There are six potential responses for each question. The participant needs to read the responses for each question carefully and indicate their level of agreement with each on a scale of 1 (Very low) to 5 (Very high).
Please help me. What method of factor analysis works well for this type of scale?
The responses to items like that would typically be treated as ordered categorical (ordinal). To run a factor analysis with ordinal items, you can use software programs (e.g., Mplus) that can properly fit factor analysis models (including both EFA and CFA) to binary and ordinal indicators based on tetrachoric/polychoric correlations using appropriate estimation methods for categorical data.
Yes, I believe the lavaan package in R can be used to run EFA with the DWLS estimator for binary/ordinal factor indicators (this estimator is referred to as WLSMV in Mplus).
Another approach which has a number of advantages over factor analysis is Mokken scaling.
While factor analysis can leave you puzzling over a table of factor loadings and trying to decide where the items belong, Mokken scaling starts with a strong definition of a univariate scale, based on the item consistency. It begins with the two items that scale together most strongly, and adds items until the resulting subscale would fail a criterion for unidimensionality. It then casts around to build another subscale from the remaining items. And so forth until the item pool is exhausted or none of the remaining items scale together.
The advantages are clear, I hope. Item subsets that are unidimensional scales, and the rejection of items that belong nowhere. Plus the availability of assumption checks to ensure that items behave well throughout their range.
There is a really excellent R package that provides all the help and functions you need to carry out the process. It is called mokken.