1: \begin{abstract}

2: The goal of this paper is to provide some tools for nonparametric

3: estimation and inference in psychological and economic experiments.

4: We consider an experimental framework in which each of $n$subjects

5: provides $T$ responses to a vector of $T$ stimuli. We propose to

6: estimate the unknown function $f$ linking stimuli to responses through

7: a nonparametric sieve estimator. We give conditions for consistency

8: when either $n$ or $T$or both diverge. The rate of convergence depends

9: upon the error covariance structure, that is allowed to differ across

10: subjects. With these results we derive the optimal divergence rate

11: of the dimension of the sieve basis with both $n$ and $T$. We provide

12: guidance about the optimal balance between the number of subjects

13: and questions in a laboratory experiment and argue that a large $n$ is

14: often better than a large $T$. We derive conditions for asymptotic

15: normality of functionals of the estimator of $T$and apply them to

16: obtain the asymptotic distribution of the Wald test when the number

17: of constraints under the null is finite and when it diverges along

18: with other asymptotic parameters. Lastly, we investigate the previous

19: properties when the conditional covariance matrix is replaced by an

20: estimator.

21: \end{abstract}

22: