diff --git a/vignettes/perform_mr.Rmd b/vignettes/perform_mr.Rmd
index 46ee8390..b8046a05 100644
--- a/vignettes/perform_mr.Rmd
+++ b/vignettes/perform_mr.Rmd
@@ -13,6 +13,7 @@ vignette: >
 ---
 
 ```{r, include=FALSE}
+evalinr44 <- getRversion() >= '4.4.0'
 knitr::opts_chunk$set(
   collapse = TRUE,
   comment = "#>",
@@ -612,7 +613,7 @@ The current plots being generated are not necessarily adequate because while the
 
 ### Using your own summary data
 
-If you want to perform analysis with your local summary data (i.e. not in the OpenGWAS database) then use then look up the `mv_extract_exposures_local()` function in replace of the `mv_extract_exposures()` function.
+If you want to perform analysis with your local summary data (i.e. not in the OpenGWAS database) then use then look up the `mv_extract_exposures_local()` function instead of the `mv_extract_exposures()` function.
 
 * * *
 
@@ -633,7 +634,7 @@ ld_mat
 
 Here `ld_matrix()` returns the LD correlation values (not R^2^) for each pair of variants present in the 1000 genomes data set.
 
-```{r eval=FALSE}
+```{r}
 dat <- harmonise_data(
   exposure_dat = bmi_exp_dat,
   outcome_dat = chd_out_dat
@@ -642,19 +643,19 @@ dat <- harmonise_data(
 
 Convert to the `MRInput` format for the MendelianRandomization package:
 
-```{r eval=FALSE}
+```{r eval=evalinr44}
 dat2 <- dat_to_MRInput(dat)
 ```
 
 This produces a list of `MRInput` objects that can be used with the MendelianRandomization functions, e.g.
 
-```{r eval=FALSE}
+```{r eval=evalinr44}
 MendelianRandomization::mr_ivw(dat2[[1]])
 ```
 
 Alternatively, convert to the `MRInput` format but also obtaining the LD matrix for the instruments
 
-```{r eval=FALSE}
+```{r eval=evalinr44}
 dat2 <- dat_to_MRInput(dat, get_correlation = TRUE)
 MendelianRandomization::mr_ivw(dat2[[1]], correl = TRUE)
 ```