On-Farm Network Research Reports

The On-Farm Network is growing, and so is access to trial results!  Single site research reports have been organized in the database below, providing detailed information at the farm level, packaged in an easy-to-read single page document.

HOW IT WORKS:

Filter the reports in the database table below by clicking the desired Crop, Year, Trial Type, and/or Major Region.  Within the database table, click on any column header to sort the table.  To view a single site report, select the Trial ID to open the single site research report in a new tab.

In the database table below, trials with significant yield differences are highlighted green in the yield difference column. On each single site report, significance is indicated by a ‘yes/no’ in the overall yield results table.

A trial that does not meet the trial requirements, eg. field history, is not included in the overall average for yield difference.

IMPORTANT INFORMATION:

There are two statistical tests that are used to analyze On-Farm Network data:

  • Paired t-tests
  • Analysis of variance (ANOVA)

The following information will help interpret results from both types of tests.

Coefficient of Variation (CV): This is the statistical measure of random variation in a trial.  The lower the value, the less variable the data.

Confidence Level: For our trials, we use a 95% confidence level. In statistics, the confidence level indicates how certain we are of the outcome of our statistical analysis.

P-value: While a confidence level tells us how certain we are of the results we get from statistical analysis, the P-value indicates if the results are statistically significant. The P-value is a probability that is calculated through the statistical analysis process. A P-value less than 0.05 indicates a statistically significant result, but a P-value greater than 0.05 indicates the results are not significant.

Interpreting Significance: So, if our statistical analysis indicates a significant yield difference, what does that actually mean? A significant yield response (where the P-value is < 0.05) means that we are 95% sure the yield difference resulted from the treatment. Alternatively, if our statistical analysis indicates there is no significant yield difference (where the P-value is > 0.05), then we are 95% certain that the treatment had no effect on yield.

Why are statistics important? Why does significance matter? Why can’t we just look at differences in yield between treated and untreated strips to determine the effect of a treatment?

Variability in yield is expected from strip to strip across an on-farm trial due to the variability that occurs across a field. So, when we get yields from each of our trial strips at the end of the season, the question is whether those yield differences are simply a result of variability in the field, or, if the yield difference is a result of the treatment/management practice investigated in the trial. We can answer that question using statistics. If the results are statistically significant, we can say that the yield difference between treatments or management practices tested in the trial was caused by the treatment or management practice. If the result is not significant, then any yield difference is likely a result of variability within the field and not a result of the treatment or management practice.

Interpreting Results – An Example: In a soybean double inoculant trial, we test the effect of double vs. single inoculant on soybean yield. Let’s say, for example, the average yield difference between double and single inoculated soybeans for one trial was 1.5 bu/ac. This yield difference will be indicated as significant or not significant. If the yield difference is statistically significant, we can say we are 95% certain that the 1.5 bu/ac increase in yield is a result of the double inoculant treatment. But, if the 1.5 bu/ac yield difference is not significant, then the double inoculant had no effect on yield compared to single inoculant and the 1.5 bu/ac yield difference simply resulted from natural variability across the trial area.

MPSG does not endorse the use of products tested in the On-Farm Network.  Although trials are conducted at multiple sites under varying conditions, your individual results may vary.  Contents of these research publications can only be reproduced with the permission of MPSG.

 

Crop

Year

Trial Type

Major Region

NOTE: To view all columns, scroll left to right

Year Region Municipality Crop Trial Type Trial Detail Yield +/- Unit Trial ID
2019 Central Portage la Prairie Dry Bean Foliar Fungicide Lance AG vs. Untreated 50 lbs/ac 2019-DBF02
2019 Central Montcalm Dry Bean Foliar Fungicide Cotegra vs. Untreated 6 lbs/ac 2019-DBF01
2018 Central Stanley Dry Bean Foliar Fungicide Cotegra vs. Untreated 1 lbs/ac 2018-DBF03
2018 Central Rhineland Dry Bean Foliar Fungicide Cotegra vs. Untreated -88 lbs/ac 2018-DBF02
2018 Central Thompson Dry Bean Foliar Fungicide Lance vs. Untreated 52 lbs/ac 2018-DBF01
2017 Central Rhineland Dry Bean Foliar Fungicide Acapela vs. Untreated 76 lbs/ac 2017-DBF01
2017 Central North Norfolk Dry Bean Foliar Fungicide Acapela vs. Lance vs. Untreated lbs/ac 2017-DBF02
2017 Central Roland Dry Bean Foliar Fungicide Lance vs. Untreated 95 lbs/ac 2017-DBF03
2017 Central Thompson Dry Bean Foliar Fungicide Lance vs. Allegro vs. Untreated lbs/ac 2017-DBF04
2017 Southwest Glenboro-South Cypress Dry Bean Foliar Fungicide Lance vs. Untreated 220 lbs/ac 2017-DBF05
2017 Central Stanley Dry Bean Foliar Fungicide Acapela vs. Untreated 14 lbs/ac 2017-DBF06
2016 Central North Norfolk Dry Bean Foliar Fungicide Lance vs. Untreated 63 lbs/ac 2016-DBF01
2016 Central Thompson Dry Bean Foliar Fungicide Lance vs. Untreated 84 lbs/ac 2016-DBF02