diff --git a/Week 3- Statistical Demand Models/multinomial_choice1.tex b/Week 3- Statistical Demand Models/multinomial_choice1.tex
index 90484dc..ad5bbec 100644
--- a/Week 3- Statistical Demand Models/multinomial_choice1.tex	
+++ b/Week 3- Statistical Demand Models/multinomial_choice1.tex	
@@ -62,7 +62,7 @@
 There are some choices that make our life easier
 \begin{itemize}
 \item Multivariate normal: $\varepsilon_i  \sim N(0,\Omega)$. $\longrightarrow$ \alert{ multinomial probit}.
-\item Gumbel/Type 1 EV: $f(\varepsilon_i) = e^{-\varepsilon_{ij}}  e^{-e^{-\varepsilon_{ij}}}  $ and $F(\varepsilon_i) = 1- e^{-e^{-\varepsilon_{ij}}}$ $\longrightarrow$ \alert{multinomial logit}
+\item Gumbel/Type 1 EV: $f(\varepsilon_j) = e^{-\varepsilon_{ij}}  e^{-e^{-\varepsilon_{ij}}}  $ and $F(\varepsilon_j) = 1- e^{-e^{-\varepsilon_{ij}}}$ $\longrightarrow$ \alert{multinomial logit}
 \item There are also heteroskedastic variants of the Type I EV/ Logit framework.
 \end{itemize}
 \end{frame}